Synthesis, Characterization, and DNA Binding Profile of a Macrocyclic β-Sheet Analogue of ARC Protein

Abstract

ARC repressor (apoptosis repressor with caspase recruitment domain) is a protein which binds selectively to a specific sequence of DNA. In humans, ARC is primarily expressed in striated muscle tissue, which normally does not undergo rapid cell turnover. This suggests that ARC may play a protective role in the prevention against Duchenne Muscular Dystrophy and several types of tumors. In this Letter we report the synthesis, characterization, and conformational analysis of a β-sheet ARC repressor mimetic, based on the amino acid sequence of the β-sheet domain in the ARC protein. The ability of this β-sheet macrocycle to bind to double-stranded DNA was also evaluated using spectroscopic methods. Our data show that the synthetic peptide has a defined conformation and is able to bind DNA with reasonable affinity. These initial results lay the groundwork for the design of novel β-sheets folded peptides as valuable substitutes of transcription factor proteins in drug therapy.

Short abstract

Macrocyclic β-sheet mimetic of ARC protein able to bind to DNA.

Gene expression is finely regulated by transcription factors (TFs). The development of synthetic entities able to mimic their ds-DNA (double stranded DNA) recognition properties could potentially cure diseases caused by alterations in their activity, such as cancer¹⁻³ and Duchenne Muscular Distrophy.⁴ ARC repressor is a transcription factor abundantly expressed in the striatum human muscle and cardiac tissue that selectively interacts with the caspase recruitment domain (CARD) of caspase-2 and caspase-8, thus inhibiting cell death caused by hypoxia and hydrogen peroxidase-mediated cells H9c2, protecting cardiac muscle from postischemic cardiomyopathy.⁵ Its abundance in long life cells suggests a protective action on the muscular fibers involved in apoptotic processes caused by mechanical stress and oxidative damage.⁵ The ARC repressor protein belongs to a superfamily of TFs called RHH (ribbon-helix-helix), on the basis of the order assumed by the elements that constitute its secondary structure (Figures 1a,1b).⁶ The crystallized complex ARC–DNA (PDB: 1PAR) reveals a tight recognition of the β-sheet (ribbon) present in the small domain of ARC to specific DNA sequences (Figure 1c).⁷ The functional unit of ARC is a dimer referred to as RHH2, characterized by a double folded symmetry and capable of binding to specific DNA sequences, including the TAGA box.^6,7 The two identical β-strands present in the amino terminal ends of each monomer form an antiparallel β-sheet (Figure 1b), which binds to the major groove of the DNA.⁸ During the recognition process the N to C direction of each strand coincides with the 3 to 5 direction of the near strand and the local β-sheet axis is bulged toward the DNA, generating the characteristic convex curvature of a two-strand that fits in the major groove (Figure 1).⁹ The β-sheet structure interacts with six consecutive base pairs by Gln, Arg, Asn side chains, establishing contacts with key sequence-specific nucleotide bases by an intensive set of hydrogen bonds.⁷ Sequence-specific binding between β-strand residues to the major groove of DNA is an exclusive feature of the RHH superfamily and distinguishes it from the ubiquitous HTH (helix-turn-helix), that binds the DNA by an α-helix.⁶ During the recognition process, the β-sheet domain undergoes an induced conformational adjustment: in the free protein, the side chains of both Phe present in each β-strands are buried in the hydrophobic core,¹¹ but when the protein binds to the DNA, these chains shift outward to fit between the oxygens of the phosphate groups, while the C^α residues of Gln and Phe move about 1.0–1.7 Å to partially fill the cavity generated by the repositioning of phenyl rings.¹² The occupation of the site operator on the DNA by ARC is governed by fine monomer–dimer equilibrium and by a cooperative oligomerization of RHH dimers when they bind to the DNA adjacent subsite operators.⁶ The analysis of the literature reveals that many efforts have been devoted to developing artificial DNA binding peptides mimicking natural TFs, but most of them rely on the use of β-helical motives of b-ZIP or Zn-Finger proteins.^13,14

Figure 1.

(a) Structure of the ARC repressor dimer in complex with its DNA target site (PDB: 1PAR). One dimer is shown in purple; the other highlights each of the monomers in green and light-blue; (b) Secondary structure of an ARC dimer. (c) Detail of the β-sheet fragment that contacts the DNA (PDB: 1PAR). The figure has been prepared by using Maestro 9.2 (Schrodinger).¹⁰ (d) Relevant CH–CH distances in Å of the β-sheet ARC repressor (PDB: 1ARR).

In contrast, the DNA binding motives based on a β-sheet like that of ARC protein have not yet been explored.

Considering the recent growing interest in the ARC protein as a therapeutic target in several human diseases,¹⁵ and the identification of the 3D structure of ARC protein in solution (PDB: 1ARR, 1ARQ)^16,17 and in complex with DNA (PDB: 1PAR),⁷ here we describe the first attempt to develop a mimetic β-sheet peptide designed to fold into a well-defined β-sheet conformation, on the basis of the structural moiety found in the native ARC protein.

The design of the first macrocyclic β-sheet ARC repressor mimetic is based on the sequence of the β-sheet portion of the native ARC protein PQFNLRWP//PQFNLRWP# (Figure 1), whereas the four prolines have been replaced by two Orn δ linkers in order to close the cycle (see product 4).

The key amino acids of each β-strand have been preserved to ensure the sequence-specific interactions with the nucleotide bases and the correct positioning of the β-sheet in the major groove.⁷ Two residues of Lys were incorporated into the native sequence to enhance the water solubility.

The Orn residues have been chosen in order to fold the molecule portions into β-hairpin turns and therefore stabilize the β-sheet organization,¹⁸ keeping the molecular symmetry and positioning necessary for the recognition of an extensive sequence of DNA bases. Macrocyclic β-sheet ARC analogue 4 was synthesized by solid phase peptide synthesis (SPPS), starting from preparing the full protected linear peptide 2 carried out on 2-chlorotrityl chloride resin by standard automated Fmoc SPPS (Scheme 1).¹⁹ Macrocyclization of peptide 2 was performed on a solution of 0.5 mM peptide concentration with 4 equiv of HCTU and 8 equiv of DIPEA in DMF over 64 h. The so obtained crude product 3 was deprotected with TFA/TIPS/H₂O 18/1/1 and lyophilized after dissolution in water, to facilitate cleavage of tryptophan carbamate (Trp-COOH) groups. The resulting crude peptide 4 was then purified by RP-HPLC and characterized by mass spectra. Lyophilization of the pure fractions yielded the macrocyclic peptide 4 in 12% overall yield (0.030 g), based on theoretic loading of the resin with Boc-Orn(Fmoc)-OH (for detailed experimental procedure, see Supporting Information). Assignments of the signals were done by ¹H NMR (Table S1) and 2D TOCSY NMR, and for the connectivity, sequential NOEs were used (see Supporting Information). In model 4 a strong NOE between CH^α Trp and CH^α Phe was found (Figure S1, Supporting Information) with the absence of the CH^α Arg-Gln NOE, which is a key feature of the β-sheet present in the ARC repressor structure. This finding, together with the sequential NOEs found (Table S2, Supporting Information), strongly suggests a quite different conformation of the model synthesized. Centered on the two Leu and Phe residues may be present a large hydrophobic cluster. The structure is also stabilized by π–π interaction between the Trp aromatic rings and lipophilic interaction between Phe aromatic rings and Leu side chains. The folding of product 4 was also studied by Circular Dichroism (CD) in 10 mM aqueous phosphate buffer (pH 7.5). Under these conditions, product 4 showed an apparently unusual CD spectrum with two intense positive bands at 195 and 225 nm (Figure S4, Supporting Information). The theoretical CD of an antiparallel β-sheet should consist of one positive band around 195 nm and one negative band around 218 nm. However, the intense exciton-coupled band between 215 and 229 nm is consistent with the interaction between indole rings, in agreement with previous NMR studies.²⁰ Considering this aspect, the CD spectrum of final product 4 fits with a β-sheet structure. DNA binding properties have been studied through fluorescence spectroscopy. Since peptide 4 contains two tryptophan amino acids, tryptophan fluorescence spectroscopy was applied to check the interaction with DNA. The tryptophan fluorescence is strongly influenced by the local environment; then its fluorescence quenching has been used to investigate protein conformational changes²¹ and protein–DNA interaction.²² In the case of DNA interaction, the quenching is mainly due to the electron transfer from indole ring to DNA bases. As expected, addition of a solution of ds-oligonucleotide DNA1, which contains the consensus target for this peptide (TAGA), over a solution of peptide 4, induces a decrease in the tryptophan signal (Figure 2a). Surprisingly, the drop of the signal almost stops completely with only 0.5 equiv of the DNA. This can be explained considering that peptide 4 might interact with the DNA in several positions simultaneously, when the concentration of DNA is lower than that of the peptide. In order to further evaluate the sequence specificity of peptide 4, the experiment was repeated by using a DNA with a double mutation in the consensus sequence (DNA2) (Figure 2b). Interestingly, the peptide showed exactly a close behavior with both DNA1 and DNA2 consensus sequences. Finally, the ethidium bromide displacement method was used to corroborate the DNA interaction, analyzing the loss of fluorescence derived from the displacement of ethidium bromide by another agent that interacts with the DNA (Figures S5,S6, SI).²³ This experiment also confirmed the interaction of the peptide with ds-DNA, and as in the case of tryptophan quenching, peptide 4 showed the same interaction with DNA1 and DNA2. Considering that peptide 4 displaces two ethidium bromide molecules after the DNA binding, we could calculate an approximate dissociation constant of 2.1 μM at 20 °C (for experimental details, see SI).²⁴ Despite the fact that the peptide does not exhibit the expected specificity, it is able to displace the ethidium bromide; this implies that it established not only electrostatic interactions with the DNA phosphates, but is docked in one of DNA grooves, making contacts with bases. Isolated transcription factor fragments typically display very low affinity for their cognate DNA sites. For example, monomeric GCN4 basic regions show a K_D of ≈ 20 μM for their target AP1 or ATF/CREB half sites,²⁵ and therefore the measured ≈2 μM binding constant shows that the designed macrocycle is capable of establishing efficient contacts with its target DNA, with affinity is in the same order of that showed by autonomous DNA-binding domains, such as the AT-Hook (K_D ≈ 6 μM),²⁶ or other small organic DNA binding agents, such as propamidine, DAPI, or dystamicin.²⁷⁻²⁹ Conceiving molecules characterized by β-sheet-like structures to address long DNA sequences is a novel and extremely challenging task, considering that the majority of proteins binding DNA and their synthetic analogues are α-helix structures capable of binding simultaneously with both the major and minor grooves of the DNA duplex.^30,31 A great amount of literature is focused on the study of ARC repressor factor as a promising target in transcription therapy.^14,32−37 In this paper we reported the first attempt to synthesize a β-sheet like macrocycle as ARC repressor analogue, in order to obtain a handle biomimetic peptide recognizing specific binding sequence of DNA. The peptide showed high affinity interaction with ds-DNA but low sequence selectivity. Based on the above observations, in order to obtain a β-sheet type peptide capable of binding to the DNA major groove, one option could consist of trying to avoid the hydrophobic collapse replacing the Leu residues with Ser, which may reduce the hydrophobic clustering and permit the desired conformation; furthermore, the substitution of the δ-linked ornithine with d-Pro-Gly sequence may also give the desired turn.³⁸ Further work is under way to improve the design process, which involves all the necessary modifications in the amino acids sequence, also tethering to minor groove binders could be a fruitful future perspective.

Scheme 1. Synthesis of the Final Product 4 as TFA Salt.

Figure 2.

DNA binding of peptide 4 studied by tryptophan fluorescence spectroscopy. (a) Fluorescence emission spectrum of a 0.5 μM solution of peptide 4 in HEPES buffer 10 mM pH 6.5, 100 mM NaCl after addition of 0, 0.5, and 1 equiv of DNA1. (b) Fluorescence emission spectrum of 0.5 μM solution of peptide 4 in HEPES buffer 10 mM pH 6.5, 100 mM NaCl after addition of 0, 0.5, and 1 equiv of DNA2. Oligonucleotide hairpin sequence (binding site in italics): DNA1: 5′-GCGAG TAGA GC TTTTT GC TCTA CTCGC-3′; DNA2: 5′-GCGAG CACA GC TTTTT GC TGTG CTCGC-3′.

Experimental Procedures

Chemicals

HPLC grade solvents were purchased from Fisher Scientific (Pittsburgh, PA); methylene chloride was further purified by passage through a column of alumina under argon. 2-Chlorotrityl chloride resin, HCTU, and all Fmoc-protected amino acids were purchased from Novabiochem (Massachusetts, USA); biochemical grade trifluoroacetic acid for HPLC was acquired from Acros Organics, and standard grade trifluoroacetic acid for deprotection of peptides was purchased from Fischer Scientific (Pittsburgh, PA). D₂O and DMSO-d₆ were acquired from Cambridge Isotopes; all other reagents were from Aldrich Chemical Co.

HPLC and Mass Spectra

Analytical reverse phase HPLC (RP-HPLC) was performed on a Beckman (Agilent Zorbax 80 SB C18 column, 50 × 4.6 mm; solvent A: H₂O/0.1% TFA, solvent B: CH₃CN/0.085% TFA). A 21.2 × 250 mm Zorbax SB-C18 PrepHT (7-μm particle size) column from Agilent on a Rainin Dynamax system was used for preparative HPLC (flow rate 10 mL/min). UV detection (254 nm) was chosen for analytical and preparative HPLC; water and acetonitrile containing 0.1% biochemical grade TFA were used as solvents. TOF-MS ESI was used to analyze all final products and reaction intermediates.

NMR Spectroscopy

Five mg of pure analogue 4 were used for the preparation of NMR samples in PP 535 NMR tubes in 0.5 mL of D₂O (2.8 mM concentration), and another 5 mg of pure analogue 4 were used for the preparation of NMR samples in PP 535 NMR tubes in 0.5 mL of D₂O 10% (0.05 mL)/H₂O 90% (0.45 mL) in 2.8 mM concentration; 5 mg of pure analogue 4 was used for the preparation of NMR samples in Shigemi tubes using 0.302 mL of D₂O (4.67 mM concentration) and another 4 mg for the preparation of NMR samples in Shigemi tubes in 0.302 mL of D₂O 10% (0.03 mL)/H₂O 90% (0.27 mL) in 3.73 mM concentration. These spectra were performed in 500 MHz Cryo Bruker and 800 MHz Varian UnityInova.

Circular Dichroism (CD) and Fluorescence measurements

CD spectra were obtained with Jasco-715 coupled with a thermostat Nestlab RTE-11, using an acquisition range: 250–190 nm; bandwidth: 2.0 nm; resolution: 0.2 nm; accumulation: 3 scans; sensitivity: 10 mdeg; response time: 0.25 s, speed: 100 nm/min. CD measurements were performed in a 2 mm cell at 20 °C. Product 4 (125 μM) was dissolved in 10 mM phosphate buffer (pH 7.5) and 100 mM of NaCl.

Tryptophan fluorescence and ethidium bromide displacement spectra were obtained with a Jobin-Yvon Fluoromax-3 (DataMax 2.20) coupled to a Wavelength Electronics LFI-3751 temperature controller, as follows: increment: 1.0 nm; integration time: 0.1 s; excitation slit width: 4.0 nm; emission slit width: 6.0 nm at 20 °C. In the case of tryptophan fluorescence, the excitation wavelength applied was 295 nm and the emission spectra were acquired from 305 to 450 nm. For ethidium bromide, the excitation wavelength applied was 540 nm and the emission spectra were acquired from 565 to 650 nm.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.5b00363.

• SPPS protocol, analytical data, conformational analysis, and binding studies of peptide 4 (PDF)

We are thankful for the support given by the Spanish grants SAF2013-41943-R, CTQ2012-31341, and CTQ2013-49317- EXP, the Xunta de Galicia GRC2013-041 and the Support of COST Action CM1105 is also kindly acknowledged. J.M. thanks the spanish government for a PhD grant.

The authors declare no competing financial interest.